Testing and evaluating CatLabs "X Film 320 Pro (2022 version)

[bernard_L]

View attachment 319007

@aparat. Kudos for performing these sytematic tests. However, I have a question re: your determination of film speed from the logE-D curves as shown on your post #53 above
Skipping other "details", the speed is based upon the point on the curve at Base+Fog+0.1. But your measured exposures do not reach low enough values to measure Base+Fog.

 

[warden]

Here's the set. Three stops either side of the "0", or 320 ISO on my Sekonic meter. Somewhere between 320 and 640 for these would be pretty nice. 200/160 is a little crispy.

View attachment 319027

Thanks for sharing that! I inverted your image in Photoshop to try to judge the lighting conditions and it helped as I have an easier time looking at positives than negatives. Anyway I think you're getting decent shadow performance of that left wall at 160 and 80. Controlling the highlights at those lower speeds would be an enjoyable experiment. Good fun, this thread.

 

[ic-racer]

Perhaps it is the graphing software, not showing the whole dataset. Which brings up the point, is the software available? I have constructed my own EXCEL spreadsheet to automatically find the parameters, but it is very much like a house of cards. One click in the wrong spot could destroy its function and take weeks to figure out. Not a fan of any Microsoft product.

 

[MattKing]

Here's the set. Three stops either side of the "0", or 320 ISO on my Sekonic meter. Somewhere between 320 and 640 for these would be pretty nice. 200/160 is a little crispy.

View attachment 319027

To me, those look over-developed. Which isn't particularly surprising, given that I expect that was what was being targeted when CatLabs arrived at with its recommendations. Negatives that I find to be under-exposed and over-developed. Those sort of negatives appear to be quite popular now.

 

[warden]

To me, those look over-developed. Which isn't particularly surprising, given that I expect that was what was being targeted when CatLabs arrived at with its recommendations. Negatives that I find to be under-exposed and over-developed. Those sort of negatives appear to be quite popular now.

OT: I "tested" a bunch of Ferrania P30 when they (like Catlabs) had no data aside from some suggested best practices. It was contrasty like Catlabs unless I exposed it at iso50 or below, and at those speeds it worked in a way that could be scanned or wet printed without too much difficulty. But for me it was an expensive time waster only to learn that the film was too slow for my intended use. That's when I got out of the business of testing undocumented films.

I offered to give the rest of my P30 to a friend that tried one of my rolls and he declined saying he was already dealing with PTSD from the first roll. ;-)

Edit: I added the quote marks around the word tested to make sure people don't think I was doing what @aparat is doing, which is on another level entirely. I don't have those skills!

 

[aparat]

I find Ilford "Relative log exposure" more intuitive as more exposure results in higher negative density. Kodak, Foma and many other film manufacturers use this convention.

Yes, I happen to agree with you. I've seen it done both ways in the literature. Please, see below for a more detailed explanation.

Could you post your software to a new thread? Some of us would appreciate being able to compute ISO, CI, etc., from densitometer measurements. I hope the software is written in a popular language with free compilers/interpreters such as Python, C, C++, or even Fortran.

Yes, I will be glad to do that. I originally started writing it to offer a free, web-based, alternative to Win Plotter, the excellent Windows application meant to be a companion to the Phil Davis BTZS textbook. That was three years ago. Then, my illness got suddenly worse and I was bed-ridden for close to three years. I started getting somewhat better earlier this summer and decided to resume working on the program. Now, I am not so sure I want to only follow the Win Plotter conventions. I think I'd like to offer alternative modes of analysis, so that's my focus now. As to the language, I am planning to use Shiny, which is a (mostly) JavaScript library meant to bring data-driven analysis apps to the web. The back-end can be written in R or (more recently) in Python. Currently, Shiny has the best compatibility and features for R, so that's what I am working on at the moment, but that may change. I still have a lot of work to do.

@aparat. Kudos for performing these sytematic tests. However, I have a question re: your determination of film speed from the logE-D curves as shown on your post #53 above
Skipping other "details", the speed is based upon the point on the curve at Base+Fog+0.1. But your measured exposures do not reach low enough values to measure Base+Fog.

Am I correct in understanding what you said about the measured exposures do not reach low enough values to measure B+F that there was too much exposure? You have a very good eye. Yes, ideally, the toes would be longer, but this was just a test run to establish how much exposure the Catlabs film needs to produce well-formed curves. There will be more analysis coming after I buy more film. The Catlabs B+F is 0.1and the exposure does produce this density on the first step of the step tablet, which is then plotted. The Arista film's B+F is 0.3, which is high. I am not sure why it's so high. I just bought this film fresh and ran the test twice to be sure. My old Arista negatives are around 0.28. The first step of the step tablet reaches the density of 0.31, so it is 0.01 higher than B+F, which is within the margin of error for my densitometer, so I just let it go. Is that what you mean?

So, the the Catlabs film, B+F is 0.1, the speed point density is 0.2, and for Arista it is 0.41. The relative log exposures for the speed points are 2.79 and 2.62, respectively. This setup is not professionally calibrated, so the results are probably only within 1/3 stop, as @Romanko pointed out. Also, please note that these numbers come from measuring the "ISO" predicted curves (curves that have the G of 0.62), not the actual curves. Hope that answers the question.

 

[aparat]

Perhaps it is the graphing software, not showing the whole dataset. Which brings up the point, is the software available? I have constructed my own EXCEL spreadsheet to automatically find the parameters, but it is very much like a house of cards. One click in the wrong spot could destroy its function and take weeks to figure out. Not a fan of any Microsoft product.

Hey, this looks fantastic! You have everything one needs to do film testing, nicely and clearly presented. I hope to make the software available at some point, but I am still far away from the finished product. There's currently no GUI at the moment. Once I have something presentable, I will share it.

 

[Oldwino]

This is a very interesting progression. Thank you for including it! It seems Rodinal 1+25 time, temperature, and agitation provided by the manufacturer work out perfectly for you. I wonder whether the 1+50 of 19 minutes gives the same contrast. Have you tried 1+50, by any chance? Also, I cannot tell from this picture, but is this a scan or a photo on a light table?

I haven't tried 1+50 yet. It would be my preferred dilution for this speed film, but 19 min seemed like a very long time, and the 9 min for 1+25 seemed more reasonable, so I started there. Once I zero in on what the film can do, I will try 1+50 next.

The pictures is an iPhone shot of the negs on a light table.

 

[Oldwino]

To me, those look over-developed. Which isn't particularly surprising, given that I expect that was what was being targeted when CatLabs arrived at with its recommendations. Negatives that I find to be under-exposed and over-developed. Those sort of negatives appear to be quite popular now.

I agree they are over developed. I just shot a half roll at 320 and reduced the development time by 10% (still with Rodinal 1+25). Results are marginally better, but very marginally. Highlights are still pretty brittle.

I am thinking that a compensating developer might help to tame this...I have some Diafine I could try.

But, I am also thinking that this is a losing battle - I think the film is just plain high contrast. There are better films if you want a wider tonal range.

I can see using this film for a particular look, in particular circumstances.

 

[Oldwino]

OT: I "tested" a bunch of Ferrania P30 when they (like Catlabs) had no data aside from some suggested best practices. It was contrasty like Catlabs unless I exposed it at iso50 or below, and at those speeds it worked in a way that could be scanned or wet printed without too much difficulty. But for me it was an expensive time waster only to learn that the film was too slow for my intended use. That's when I got out of the business of testing undocumented films.

I offered to give the rest of my P30 to a friend that tried one of my rolls and he declined saying he was already dealing with PTSD from the first roll. ;-)

Edit: I added the quote marks around the word tested to make sure people don't think I was doing what @aparat is doing, which is on another level entirely. I don't have those skills!

Ha Ha Ha! That's exactly my experience with P30. I gave my rolls away, too.
Hopefully, I can tame the Cat.

 

[MattKing]

But, I am also thinking that this is a losing battle - I think the film is just plain high contrast. There are better films if you want a wider tonal range.

I'd still like to see more results with a meaningful amount of more exposure, plus less development.

 

[aparat]

I'd still like to see more results with a meaningful amount of more exposure, plus less development.

Matt, I agree and I am working on it. I should have a family of curves ready sometime next weekend. They will be based on what seems, empirically, to be the "correct" amount of exposure for standard film testing. I will also try to take a few photographs to go along with the data.

 

[MattKing]

I look forward to both.

 

[bernard_L]

Am I correct in understanding what you said about the measured exposures do not reach low enough values to measure B+F that there was too much exposure? You have a very good eye. Yes, ideally, the toes would be longer, but this was just a test run to establish how much exposure the Catlabs film needs to produce well-formed curves. There will be more analysis coming after I buy more film. The Catlabs B+F is 0.1and the exposure does produce this density on the first step of the step tablet, which is then plotted. The Arista film's B+F is 0.3, which is high. I am not sure why it's so high. I just bought this film fresh and ran the test twice to be sure. My old Arista negatives are around 0.28. The first step of the step tablet reaches the density of 0.31, so it is 0.01 higher than B+F, which is within the margin of error for my densitometer, so I just let it go. Is that what you mean?

So, the the Catlabs film, B+F is 0.1, the speed point density is 0.2, and for Arista it is 0.41. The relative log exposures for the speed points are 2.79 and 2.62, respectively. This setup is not professionally calibrated, so the results are probably only within 1/3 stop, as @Romanko pointed out. Also, please note that these numbers come from measuring the "ISO" predicted curves (curves that have the G of 0.62), not the actual curves. Hope that answers the question.

After posting my comment #76, I thought maybe you had measured B+F outside the step wedge, in an area totally not exposed (which is the correct way). Or, as suggested by ic-racer in post #78, "Perhaps it is the graphing software, not showing the whole dataset."

But the numbers that you state above (blue text) show that you just take as B+F the lowest exposure point in the step wedge. Most obvious is "The Arista film's B+F is 0.3" (blue curve). I find it hard to believe that the blue curve, if measured further to the left, would not reach a density level of 0.2. Can you please, if you have not already done so, measure B+F in a film area that was totally un-exposed?
If true, that would mean that the Arista film's speed point is further to the left, i.e. its speed is higher; and, even if the absolute photometric calibration is not correct, that would also mean that the Catlab film is even slower relative to a "ISO100" film.

This said:

• Leave aside 1/3 stop quibbles; in b/w that is completely insignificant; in actual use much below the inevitable differences between pointing your meter a little more to the right or to the left.
• The test by oldwino https://www.photrio.com/forum/threa...film-320-pro-2022-version.194630/post-2599132 tends to show rather convincingly that 320 is not an unreasonable speed for that film. Could the long toe explain this apparent contradiction?

 

[AgX]

@aparat. But your measured exposures do not reach low enough values to measure Base+Fog.

They do not need to be, if for instance one has established already the B+F-density in a seperate test, outside a sensitometer. It then would be sufficient to have graded exposures above that level and below the qualification level.

 

[bernard_L]

They do not need to be, if for instance one has established already the B+F-density in a seperate test, outside a sensitometer. It then would be sufficient to have graded exposures above that level and below the qualification level.

Did you read the whole of my post?

maybe you had measured B+F outside the step wedge, in an area totally not exposed (which is the correct way)

Did you read this from aparat:

Am I correct in understanding what you said about the measured exposures do not reach low enough values to measure B+F that there was too much exposure? You have a very good eye. Yes, ideally, the toes would be longer

Implying (pending confirmation by aparat himself) that he derived the B+F from the curve(s) as displayed, at the leftmost point.

Do you really believe that the blue curve (Arista) if extended to the left (by proper measurements as outlined above) would not go below D=0.3? Hardly any sign of the toe region beginning.

 

[aparat]

After posting my comment #76, I thought maybe you had measured B+F outside the step wedge, in an area totally not exposed (which is the correct way). Or, as suggested by ic-racer in post #78, "Perhaps it is the graphing software, not showing the whole dataset."

But the numbers that you state above (blue text) show that you just take as B+F the lowest exposure point in the step wedge. Most obvious is "The Arista film's B+F is 0.3" (blue curve). I find it hard to believe that the blue curve, if measured further to the left, would not reach a density level of 0.2. Can you please, if you have not already done so, measure B+F in a film area that was totally un-exposed?
If true, that would mean that the Arista film's speed point is further to the left, i.e. its speed is higher; and, even if the absolute photometric calibration is not correct, that would also mean that the Catlab film is even slower relative to a "ISO100" film.

This said:

• Leave aside 1/3 stop quibbles; in b/w that is completely insignificant; in actual use much below the inevitable differences between pointing your meter a little more to the right or to the left.
• The test by oldwino https://www.photrio.com/forum/threa...film-320-pro-2022-version.194630/post-2599132 tends to show rather convincingly that 320 is not an unreasonable speed for that film. Could the long toe explain this apparent contradiction?

View attachment 319213

Thanks for your questions! It's great to have them, they bring more detail into the thread. Let me see if I can answer your questions. Yes, the entire data set is plotted here. And, yes, I always measure B+F on a clear part of the film. My "sensitometer" has a piece of opaque plastic off to the side of the step tablet to make sure it is not exposed. See the image below, clear film patch marked with an arrow.


The Arista film, as I mentioned earlier, has a very high B+F, of about 0.31 - 0.32, subject to the error of my sensitometer. So, the toe is very short. The first step of the step tablet does not reach B+F by approximately 0.01, which I ignored because its within the margin of error. Simply put, the Arista film received too much exposure to produce a well-formed toe. However, the Arista is only meant to provide a point of reference for the Catlabs film exposure. That was the whole point, namely, to establish the "right" amount of exposure for the Catlabs film and to compare it another popular film, which I'd tested before and whose performance I am familiar with. I concluded from this initial test, that I need to reduce exposure by about 1/2-2/3 stop to get a long enough toe for precise analysis.

And, yes, you are correct, the Catlabs Pro 320 does, indeed appear slower than the Arista Edu Ultra 100 by about 1/2 stop.

This may have gotten lost in this thread, but I ran the original test using a calibrated and certified sensitometer, giving the Catlabs film the "correct" amount of exposure for an ISO 200-320 film. I ran that full test twice, and spot-tested the film the third time, to be sure. I also compared those results to Neopan 400 and 400TX. The Catlabs film turned out to be much slower than either. You are absolutely correct that the long toe makes curve analysis very difficult and somewhat unreliable. However, I could not use my sensitometer to test the film further, as it does not allow for the change in illumination levels, only in duration. I would have to expose the Catlabs film for a calculated exposure of 2-3 seconds. Without knowing reciprocity data from the manufacturer, that would have been futile.So I constructed a DIY sensitometer to give the film a lot more exposure, this time for the shutter duration of 0.25 sec., so there's no contribution from reciprocity failure. The curve you're referring to, was obtained with that DIY device. I am including a more detailed plot of the Catlabs curve, so you can see more easily some of the important parameters.


How do I explain the apparent discrepancy between my results and those of others? I can't without running the full test yet again, this time, with the "correct" amount of exposure, such that we get a well-formed curve family. I am currently working on that, but it will take me at least a week or two to complete. I might also try to do a rudimentary spectral response analysis.

To sum up, in my tests, the Catlabs Pro 320 is consistently slower than both Kodak 400TX and Arista Edu Ultra 100. Once my test has been concluded, I will make available a full report, including multiple methods of measuring film speed and the raw data, so that anyone here can run their own analysis, with their own tools. Then, and only then, could we reach any sort of conclusion as to what the Catlab Pro 320 film speed is, what its performance is, in general, and response to development time, in particular.

 

[bernard_L]

@aparat. Thank you for your patience with my questions. My apologies for incorrectly assuming that you had taken a shortcut to the determination of B+F.
Again congratulations for your systematic measurements. If you are interested in similar topics, you might like this book:
(see buyers comments)
although at the current price I cannot recommend to buy it (I got mine 5 years ago for 10€+shipping!!) Maybe borrow from a public library.
Final note. Is that film, most probably a rebranding of a known film from a major brand, worth your efforts?

 

[aparat]

I hope this is not off topic. If it is, perhaps the moderator can move it to another thread or delete it. I've been away from Photrio for a few years. Before I left, I thought, perhaps erroneously, that, in the Photrio community, the "gold standard" for evaluating film performance was in terms of producing an excellent silver gelatin (or alternative process) print of an average scene, in a straightforward manner (i.e., without film masking, etc.). So, film speed, for example, would have to be considered within that context, i.e., to obtain minimum exposure required for a beautiful print. Sensitometric testing is meant to give us an generalizable, empirical account of film performance with regard to that goal. After all, we cannot test each film by doing a proper psychometric study of print quality judgments across a sizable, randomly sampled population. Is that still the broadly held belief around here? I learned a lot from Photrio members over the years!

Why do I ask? Well, I guess a lot has changed over the past few years. There are a lot of new films available, some with unconventional performance characteristics, some many years expired, some meant for aerial surveillance. There's a new generation of film photographers who aren't interested in data sheets, who don't print in the darkroom at all. They shoot, process, and digitize the film, most commonly, it seems, with a modern DSLR, and share their images online. What I gathered from the original Catlabs thread was that the company is marketing the film to that kind of customer base. There's absolutely nothing wrong with that, as far as I am concerned. To be honest, I wish I had the equipment and skills to do that myself.

However, that kind of film use might require a different set of diagnostic criteria. Film holds a lot of data. A modern digital camera can extract a lot of that data, and RAW processing software, such as Negative Lab Pro, Adobe Lightroom, Photoshop, etc., can make impressive use of all that information and produce beautiful images with ease, even from negatives that would be otherwise considered imperfect. So how do we test film in this new reality? Should we still care about sensitometry at all? How do we reconcile sensitometric results with on-screen evaluations of digital images?

 

[BrianShaw]

All good questions and observations. I would only suggest that the production of “fine photography prints” and the quest for perfection and the desire for scientific detailed data never was a quest by “everyone” on APUG. I, for instance, respect all of that but just want to make decent images of interesting subject matter that are well composed, and if they are technically flawed… so be it. I know that I’m not alone in preference for an artistic approach to photography than to quest for perfection, sometimes it seems at the expense of doing photography at all.

I’m thrilled that you are doing what you are doing. I’m equally thrilled by the work done by folks who have put the film in their cameras and taken pictures using the vendor-provided information and guidance. It all really is value added for both kinds of photography.

 

[Dwight Anderson]

Based on what I have seen from the photo meetup group I joined this summer, the photographers in their 20's and 30's are very interested in unusual films. They grew up with digital cameras that are too sterile and predictable, so they view film as a way to experiment with unpredictable results. Few of them have an enlarger and if they do want a print, they take it to a lab where it's printed digitally. Personally, I'm not a big fan of unusual films, but I have to admit the people in this group take wonderful photos with films like Lomo purple, Cinestill 800 and others that don't perform like Ektar or Tmax. Our last photo-walk was sponsored by Lomo Metropolis, who gave out some free film and previewed their new point and shoot 35mm camera with a 21mm lens, built in color filters for the flash and slide on attachments like a kaleidoscope, close up lens and frame splitter. It'll probably sell very well.

What Aparat is doing is very useful and I like the straightforward way he is presenting it, but I think many film buyers today are using a hybrid process that allows them to create the look they want to present on screen, regardless of the film characteristics we might find important.

 

[aparat]

All good questions and observations. I would only suggest that the production of “fine photography prints” and the quest for perfection and the desire for scientific detailed data never was a quest by “everyone” on APUG. I, for instance, respect all of that but just want to make decent images of interesting subject matter that are well composed, and if they are technically flawed… so be it. I know that I’m not alone in preference for an artistic approach to photography than to quest for perfection, sometimes it seems at the expense of doing photography at all.

I’m thrilled that you are doing what you are doing. I’m equally thrilled by the work done by folks who have put the film in their cameras and taken pictures using the vendor-provided information and guidance. It all really is value added for both kinds of photography.

Thank you! I didn't mean to imply that the younger crowd do not seek perfection, only that traditional sensitometric analysis may not be applicable/useful to their process and their pursuit of photographic perfection.

 

[aparat]

@aparat. Thank you for your patience with my questions. My apologies for incorrectly assuming that you had taken a shortcut to the determination of B+F.
Again congratulations for your systematic measurements. If you are interested in similar topics, you might like this book:
(see buyers comments)
although at the current price I cannot recommend to buy it (I got mine 5 years ago for 10€+shipping!!) Maybe borrow from a public library.
Final note. Is that film, most probably a rebranding of a known film from a major brand, worth your efforts?

Thanks so much for keeping me on my toes, and for the book recommendation! I don't have it, so I'll try to find it as soon as I can. As to the effort, once I started, I'd like to finish , plus one can always learn something new from forum members.

 

[Oldwino]

I hope this is not off topic. If it is, perhaps the moderator can move it to another thread or delete it. I've been away from Photrio for a few years. Before I left, I thought, perhaps erroneously, that, in the Photrio community, the "gold standard" for evaluating film performance was in terms of producing an excellent silver gelatin (or alternative process) print of an average scene, in a straightforward manner (i.e., without film masking, etc.). So, film speed, for example, would have to be considered within that context, i.e., to obtain minimum exposure required for a beautiful print. Sensitometric testing is meant to give us an generalizable, empirical account of film performance with regard to that goal. After all, we cannot test each film by doing a proper psychometric study of print quality judgments across a sizable, randomly sampled population. Is that still the broadly held belief around here? I learned a lot from Photrio members over the years!

Why do I ask? Well, I guess a lot has changed over the past few years. There are a lot of new films available, some with unconventional performance characteristics, some many years expired, some meant for aerial surveillance. There's a new generation of film photographers who aren't interested in data sheets, who don't print in the darkroom at all. They shoot, process, and digitize the film, most commonly, it seems, with a modern DSLR, and share their images online. What I gathered from the original Catlabs thread was that the company is marketing the film to that kind of customer base. There's absolutely nothing wrong with that, as far as I am concerned. To be honest, I wish I had the equipment and skills to do that myself.

However, that kind of film use might require a different set of diagnostic criteria. Film holds a lot of data. A modern digital camera can extract a lot of that data, and RAW processing software, such as Negative Lab Pro, Adobe Lightroom, Photoshop, etc., can make impressive use of all that information and produce beautiful images with ease, even from negatives that would be otherwise considered imperfect. So how do we test film in this new reality? Should we still care about sensitometry at all? How do we reconcile sensitometric results with on-screen evaluations of digital images?

I think a lot of the new films we are seeing, whether actual new emulsions or simply rediscovered old ones, come about because of the shift in how people work with images, and expect images to look. The decline of the darkroom and the ascendence of digital photo manipulation, plus the general look of digital photography (higher native contrast, sharpness and acuity), mean that there is more room now for what many would deem "experimental" films. The overall look of CatLabs 320 is one example - it has an inherent graphic, high contrast look. It is not meant to faithfully record a scene (that requirement is met by a phone camera nowadays), but rather to convey an emotion.
These film are popular because they look "sort-of" digital (they have a familiar contrast structure), and they have a built-in "filter". Part of the charm and attraction may be that this "filter" (look) is somehow predetermined by the film and not by the users choice - in other words, you can have a cool looking shot "straight out of the camera" instead of with an Instagram filter. That makes it seem more legitimate to some.

I remember as a kid (a while ago now) discovering Tech Pan and other high contrast copy films, and marveling over how different you could make the world look.

There's a line in a recent Nick Cave song which says "we're sick and tired of seeing the way things are". Reality is sometimes boring - who needs to look at photos with lots of shadow detail and perfect highlights? At least that's the impression I get.

 

[MattKing]

However, that kind of film use might require a different set of diagnostic criteria.

You may very well be correct - or at least it may be the case that additional criteria may be useful.
In my experience though, the old criteria that arose during darkroom print days is just as informative and useful in an environment that requires the digitization of film.
The "old" criteria helps photographers understand how to get useful stuff on to the film. The new digital tools just give those photographers more options on what can be done with it.
This is simply a guess, but with respect to this "new" film, it wouldn't surprise me to eventually learn that it will be useful as a lower speed (= less grain) version of the 3200 labelled films. I don't know why that sort of flexibility isn't trumpeted as a marketing advantage, rather than avoided at all costs.